CANCELLED: Optimization of Node-level Performance and Scaling on Hawk (HPE Apollo)

Research & Science Enterprises & SME Optimization & Scalability
CANCELLED: Optimization of Node-level Performance and Scaling on Hawk (HPE Apollo)

Overview

June 22: This course has been cancelled.

May 14: This course will be provided as ONLINE course (using Zoom).

In this workshop, users can apply the know-how gained from the node-level performance engineering course (which takes place from Monday to Wednesday of the very same week) while being assisted by HLRS, HPE and AMD staff as well as one of the instructors of the node-level performance engineering course. By means of this assistance, it's not only possible to enhance the node-level performance but also the scaling of the codes. By doing so, users can raise the quantity and/or quality of their scientific findings while the costs (in terms of core hours) remain constant.

In order to achieve usable performance enhancements, it is important to discuss pros and cons of potential solutions. This, however, requires application as well as machine expertise. Hence, this workshop brings together our users (with their application expertise) and support staff (HLRS, HPE and AMD staff as well as an instructor of the node-level performance engineering course) with their machine expertise.

Target audience: Groups holding a compute time budget to be used on Hawk.

Agenda & Content

Supported by HLRS, HPE and AMD specialists as well as an instructor of the node-level performance engineering course, you will enhance the node-level performance and scaling of your code. In order to do so, you will analyze the runtime behavior, locate bottlenecks, design and discuss potential solutions as well as implement them. All categories of bottlenecks (CPU, memory subsystem, communication and I/O) can be addressed, according to the requirements of the code.

First day:

08:45-09:00 Registration 
09:00-09:15 Introduction   
09:15-17:30 Optimization of your own applications on Hawk

Second day:

09:00-15:30 dto.

Important: If you plan to join the workshop, we recommend

Prerequisites
Prerequisites
  • Your group holds a compute time budget to be used on Hawk.
  • In order to attend the workshop, you should already have an account on Hazel Hen and your application should already be running on the system.
  • Please set up a test case to be used. While doing so, please adhere to the following requirements:
    • number of cores:
      • In order to be representative, the test case should be in size comparable to your current and maybe future production runs.
      • In order to save valuable resources, it should however be as small as possible.
      • So take into account to reduce the simulated domain and keep the computational load per node constant ("weak down scaling"). Ideally, the test case should use about 64 nodes.
    •  wall time:
      • In order to allow for a productive as well as effective workflow, the wall time should be a few minutes only.
      • At the same time, it should cover all important parts of the code, including I/O.
      • So take into account to reduce the number of simulated time steps.
      • Please figure out the time-stepping loop in your code so that analyses can be focused on that part. 
Language
Language

The course language is English.

Course material
Course material

See recordings of https://www.hlrs.de/training/2020/HY-S.

Teacher
Teachers

HLRS, HPE and AMD staff as well as an instructor of the node-level performance engineering course taking place on the days before this workshop

Registration

via online registration form

Extended Deadline
Deadline

for registration is June 14, 2020 (extended deadline)

Late registrations after the deadline are still possible but maybe with reduced quality of the handouts.

Fee
Fee

Students without Diploma/Master: 0 EUR
Students with Diploma/Master (PhD students) at German universities: 0 EUR
Members of German universities and public research institutes: 0 EUR
Members of universities and public research institutes within EU or PRACE member countries: 0 EUR.
Members of other universities and public research institutes: 60 EUR
Others: 300 EUR

(includes coffee breaks)

Organization

Travel Information and Accommodation

see our How to find us page.

PRACE PATC and bwHPC-C5

HLRS is part of the Gauss Centre for Supercomputing (GCS), which is one of the six PRACE Advanced Training Centres (PATCs) that started in Feb. 2012.
HLRS is also member of the Baden-Württemberg initiative bwHPC-C5.
This course is provided within the framework of the bwHPC-C5 user support. This course is not part of the PATC curriculum and is not sponsored by the PATC program.

Local Organizer

Rolf Rabenseifner phone 0711 685 65530, rabenseifner@hlrs.de
Khatuna Kakhiani phone 0711 685 65796, kakhiani@hlrs.de
Björn Dick phone 0711 685 87189, dick hlrs.de

Shortcut-URL & Course Number

https://www.hlrs.de/training/2020/HPE2

Further Courses

http://www.hlrs.de/training/ and http://www.hlrs.de/training/overview/ (at HLRS)

See also the Nodel-Level Performance Engineering course 2020/NLP. One of the NLP trainers will stay for the first day of this scaling workshop.